Twenty-five years after the landmark patient safety report "To Err is Human" was published, only marginal improvements have been realized.1 However, recent policy efforts have emerged to transform patient safety, including recommendations from the President鈥檚 Council of Advisors on Science and Technology, a new structural measure of patient safety from the Centers for Medicare & Medicaid Services (CMS), and bipartisan support for a National Patient Safety Board.2,3 As health care enters an artificial intelligence (AI) revolution, these policy efforts should sift hype from reality to best use health information technology (IT) and AI. A major focus should be on how health IT and AI can improve patient safety gaps while mitigating unintended consequences.